有了Gabor核函数后就可以采用前文中提到的“离散二维叠加和卷积”或“快速傅立叶变换卷积”的方法求解Gabor变换，并对变换结果求均值和方差作为提取的特征。32个Gabor核函数对应32次变换可以提取64个特征（包括均值和方差）。由于整个变换过程代码比较复杂，这里仅提供测试代码供下载。该代码仅计算了一个101×101尺寸的Gabor函数变换，得到均值和方差。代码采用两种卷积计算方式，从结果中可以看出，快速傅立叶变换卷积的效率是离散二维叠加和卷积的近50倍。-to extract features. 32 Gabor kernel function corresponding to 32 times transform can extract 64 characteristics (including mean and variance). Because the whole transformation process code is complicated, only provide test code for download here. This code can only calculate a 101 x 101 size of the Gabor function transform, get the mean and variance. Code USES two kinds of convolution calculation, can be seen　the results, the efficiency of the fast Fourier transform (convolution is discrete two-dimensional superposition and nearly 50 times of convolution.